Part and parcel of any technology initiative is the task of demonstrating a clear and measurable return on investment (ROI). AI and generative AI (genAI) take this already formidable challenge to the next level. As many are learning, the technologies don’t map as well to traditional financial-oriented metrics, but rather to less tangible benefits that are harder to quantify.
The lack of standardized metrics, difficulties aligning AI initiatives with core business objectives, and an emphasis on longer-term “soft returns” all contribute to this uphill battle, particularly around AI-driven customer experience (CX) use cases. Yet even without empirical evidence, companies are already placing bets that AI will be central to reshaping contact centers and customer interactions. The Everest Group projects AI will drive nearly $100 billion in value across the customer service market by 2027.[1]
Foundry reached out to the CIO Experts Network, a community of IT professionals and technology industry influencers, to explore what’s different about ROAI (return on AI investment) compared to traditional technology initiatives. We also gathered best practices to help secure both short- and long-term buy-in.
Beyond productivity and cost reductions
According to IDC, for every $1 a company invests in genAI, the ROI is 3.7x, and the most advanced companies can realize up to 10x. However, many of the short-term benefits are keyed to productivity improvements, which can be a trap. Many C-suite leaders only look for AI investments to reduce costs and drive efficiencies, according to Issac Sacolick (@nyike), president of StarCIO and a bestselling author.
It’s important to take a sharper lens when considering ROAI, including how quickly a system learns, how much it compresses time-to-insights, and how it aligns cross-functional teams to mor quickly make better decisions. Short-term wins like fewer iterations or reduced testing costs can help build confidence in AI, but the long-term value comes from reorienting the entire organization to be more adaptive and signal driven, says Kumar Srivastava, chief technology officer at Turing Labs.
“If we’re only measuring dollars, we’re missing the real story,” Srivastava says. “This is about unlocking smarter execution at scale.”
Establish short-, mid-, and long-term goals
To capture ROAI, organizations must move beyond dollars and headcounts and refocus the measurement lens on operational efficiency, decision velocity, and customer satisfaction, influencers said.
Start by identifying which manual workflows or decision points AI enhances and then track metrics such as time-to-resolution, employee satisfaction, and model accuracy over time. Along the way, it’s important to frame benefits in familiar business terms. “Saying things like enabling faster market entry or reducing compliance risks makes the case for sustained investment,” says Will Kelly, a writer focused on AI and cloud technologies.
Any kind of improvement also starts with a baseline, and that same concept applies to quantifying ROAI. Cross-department collaboration and input from stakeholders at every level can help establish a baseline and formalize acceptable ROAI metrics, says Arsalan Khan (@ArsalanAKhan), speaker, advisor, and blogger.
Short-terms gains are generally straightforward. Allstate, for example, determined that AI-assisted customer support has delivered a 14% increase in issues resolved per hour (NBER), which showcases tangible productivity boosts, according to William Benjamin, principal generative AI/ML expert at the insurance giant. Metrics like reduced operational costs, faster response times, and higher sales are more concrete evidence of AI’s value, easily validated through controlled experiments, he says.
Longer term, Allstate is evaluating things like customer lifetime value (CLTV) and retention—critical metrics that are arguably harder to quantify. “This is where causal inference and predictive modeling become critical,” Benjamin says. “By using leading indicators such as reduced claim handling time to reliably forecast strategic outcomes like CLTV, we can quantify the downstream impact of AI and build a data-backed case for sustained investment.”
There are some common areas where companies can cultivate mid- and longer-term value from AI. In the mid-term, companies can look to AI to deliver value for improving customer satisfaction (CSAT) scores, bolstering marketing lead conversions, and increasing organizational intelligence through deployment of large language models (LLMs) and AI agents into employee workflows, says StarCIO’s Sacolick. Longer-term value will come through revolutionizing product interfaces by exposing AI agents directly to customers.
“For long-term metrics, C-level leaders should consider the impact on revenue, customer retention, and the disruptions that occur when competitors become AI-first businesses, and customers shift to agentic product and service offerings,” Sacolick explains.
As organizations build AI maturity, they will need to revisit and evolve metrics as part of their on-going journey. “Early on, focus on wins like faster time to market and cost avoidance,” adds Michael Bertha, a partner at Metis Strategy. “These initial gains may not hit the P&L directly, but over time, they compound into scalable cost savings that do.”
As always, security needs to be a high priority and is crucial for achieving ROAI. Organizations need to mitigate any potential risks associated with their AI models and machine learning environments to avoid potential problems that can undermine results. “If these systems are not properly secured, attackers could exploit vulnerabilities to exfiltrate sensitive data, compromise AI models, or poison data sets,” cautions Aaron Momin, chief information security officer at Synechron.
The bottom line
In the end, ROAI requires a different mindset, requiring stakeholders to pivot away from measuring what AI delivered to assessing if an organization is ready to deliver with AI.
As part of the shift, transition away from justification of spending to ensuring success through capability building. Similarly, move beyond evaluating the performance of existing AI to validating readiness for future AI efforts. “To unlock real AI value, shift your mindset from measuring outcomes to enabling them,” says Peter Nichol, data and analytics leader for North America at Nestlé Health Science. “ROI doesn’t start with algorithms; it starts with readiness.”
[1] Everest Global, Generative AI and a Holistic CX Platform: A Perfect Pairing, 2025.